The strategy of targeting transcription using small molecule therapeutic agents that bind to specific RNAs can be applied to the field of antivirals. Human immunodeficiency virus (HIV) is the infectious agent responsible for acquired immunodeficiency syndrome (AIDS). Replication of the virus is dependent on specific interactions between viral ribonucleic acids (RNAs) and viral proteins. For example, one such crucial interaction is between the protein Rev and the virally encoded RNA sequence termed the Rev responsive element (RRE). The REV/RRE interaction is deemed to be critical for viral replication and has been described as the "Achilles' heel" of the virus. Dr. F. Wong-Staal has said that the REV/RRE interaction represents "the best target" for anti-HIV drug design. (Cohen, J., Science, 1993, 260, 1257.) The importance of the REV/RRE interaction has been established by two methods: 1) Using antisense directed at the Rev mRNA, Matsukura and coworkers (Matsukura, M., et al., Proc. Natl. Acad. Sci. USA, 1989, 86, 4244-4248) decreased the expression of the Rev protein; and 2) Expression of dominant negative mutants of the viral regulatory protein abolishes Rev activity in a cotransfection assay. (Malim, M. H., et al. Cell, 1989, 58, 205-214.; Bevec, D., et al. Proc. Natl. Acad. Sci. USA, 1992, 89, 9870-9874.) However, both methods are severely limitated as therapeutic approaches as they require delivery of very large biomolecules to the infected cells of the patient.
In order to address HIV resistance it is likely that new therapeutic approaches will require combination therapy. Inhibition of the REV/RRE interaction with small orally active molecules will be very appropriate for such combinations because their mode of action and locus of activity are quite distinct from reverse transcriptase and protease inhibitors.
Experiments from the laboratories of Zapp and coworkers (Zapp, M., et al. Cell, 1993, 74, 969-978) have demonstrated, for the first time, that a small molecule, the aminoglycoside antibiotic neomycin B, could disrupt the REV/RRE interaction. This work of Dr. Michael Green and coworkers at the Howard Hughes Medical Institute, University of Massachusetts Medical Center has demonstrated that a limited structure activity relationship (SAR) for the inhibition of the REV/RRE interaction can be based on bio-activity data obtained from commercially available aminoglycoside antibiotics and from neomycin analogs that have been prepared. However, the structures of the aminoglycoside antibiotics are synthetically challenging and do not lend themselves to the rapid preparation necessary for a medicinal chemistry program. RNA targeted approaches are similarly suited to many other viral targets, such as Hepatitus B and C.
Ribonucleic acid plays a central role in the life of organisms from bacteria to humans. RNA serves as the intermediary between DNA and protein, the final product of gene expression and another primary building block of all living things. DNA is transcribed into a complementary RNA strand carrying the protein blueprint. The RNA interacts with ribosomes and is translated into protein. Each RNA can be translated into many copies of protein, so a small amount of RNA leads to a large amount of protein.
However, RNA is not just a transient messenger in this process. RNA directly controls gene expression by catalyzing the processing and translation of messenger RNA into protein. Ribosomes are massive ribonucleotide complexes, involving multiple protein-RNA interactions, in which RNA forms the core catalytic component. RNAse P is a smaller ribonucleotide complex that is critical for the processing of multiple RNAs such as tRNAs and 4.5S rRNA. Furthermore, RNA has been shown to play a critical role in regulation of the synthesis of ribosomal proteins from at least 5 different operons encoding ribosomal proteins and RNA polymerase subunits. Thus, RNA not only plays a critical role in the translation process but also regulates the synthesis of the translation machinery. RNA is clearly a central player in the regulation of the most fundamental processes of the cell.
Interest in targeting RNA emerged in the mid-1980's with the discovery of antisense oligonucleotides. These antisense oligonucleotides are large molecule, synthetic fragments that are designed to bind to RNA for the purpose of inhibiting or regulating its activity. While the approach of using antisense oligonucleotides is promising, to date the technology has not yielded any approved drugs. Two critical challenges in the development of antisense oligonucleotides remain: 1) effective delivery of these macromolecules into the cell, and 2) the development of new technology for the appropriate manufacture of large quantities of material.
The design of antisense molecules is based on the assumption that RNA is a linear, unfolded template without structure. The unexpected discovery by Cech and Altman (McClure, W. R., et al., J. Biol. Chem., 1978, 253, 8949-8956.; Altman, S., et al., FASEB J., 1993, 7, 7-14.) of catalytic RNA, however, has forever changed this view. Their studies sparked a flurry of investigations into the structure and function of RNA that have led to unexpected demonstrations of the enormous plasticity of RNA.
Despite the existence of tRNA x-ray crystal structures, the compact, stable structure of tRNA was not considered representative of other, larger RNA species. It was not until the advent of nuclear magnetic resonance that scientists were able to determine and appreciate the complexity, rigidity and diversity of RNA structures in general. This diversity endows RNA with its extraordinary functional versatility. RNA exists in both single stranded and helical duplex forms. These duplexes are distorted by loops, bulges, base triples and junctions between helices. The structural diversity of RNA is far greater than that of DNA, and similar to that of proteins, making RNA a suitable target for small molecule drugs. Facile methods for crystallizing RNA have now been developed, which will greatly advance the understanding of RNA structure and function (Doudna et al., Proc. Natl. Acad. Sci. USA, 1993, 90, 7829-7833).
These recent advances in understanding the function and structure of RNA have provided a foundation for discovery and development of small molecule compounds that bind to and regulate RNA function. Small molecules have important advantages over antisense oligonucleotides and currently represent the vast majority of therapeutic compounds. Many small molecule drugs are active for oral administration and can be produced through conventional chemical synthesis techniques.
Small molecules can bind RNA and block essential functions of the bound RNA. Examples of such molecules include erythromycin, which binds to bacterial rRNA and releases peptidyl-tRNA and mRNA, and aminoglycoside antibiotics, which bind to the decoding region of 16S rRNA and cause a misreading of mRNA codons during translation. Both genetic and biochemical experiments confirm the specificity of these interactions. Also, unequivocal evidence supports the binding of thiostrepton to a specific site in 23S rRNA. Bacterial resistance to kanamycin and gentamicin maps to the specific methylation of 16S rRNA. Aminoglycosides have been chemically footprinted on 16S RNA in the absence of ribosomal proteins. Green and co-workers have shown that 2-DOS-containing aminoglycosides bind specifically to the HIV RRE, block binding of the HIV Rev protein to this RNA, and inhibit HIV replication in tissue culture cells (Zapp, Stern and Green, 1993). (Zapp, M. et al. Cell, 1993, 74, 969-978.)
The specificity and affinity of small molecule ligands of specific RNAs have also been demonstrated by the selection of aptamers, RNAs that have been selected from populations of random sequence RNA molecules for affinity to a specific, small molecule ligand. Many of these aptamers bind with high affinity to their cognate small molecule ligand. Examples of small molecule aptamers include RNAs with high affinity for the dyes Cibacron Blue 3G-A and Reactive Blue 4, cyanocobalamin, and theophylline. The ability of RNA to distinguish between two closely related molecules is best demonstrated by the ability of the theophylline aptamer to discriminate between theophylline and caffeine, which differ by only a single methyl group.
In light of these advances, it is apparent that targeting RNA (mRNA, tRNA, rRNA, etc.) should provide effective methods for the design of new therapeutics.
Inhibition of RNA polymerase in bacteria is one of the many enzyme inhibition targets that can be addressed by screening a universal combinatoral library. A single, multisubunit enzyme is responsible for all RNA synthesis in bacteria. RNA polymerase is a good target for antimicrobial therapy, because transcription is an essential process in the bacterial life cycle. Thus, this approach also has the advantage of allowing the selection of agents that will have a high probability of being bactericidal rather than bacteriostatic, in contrast to current cell-based assays, which do not distinguish between bacteriostatic and bactericidal compounds. Because the bacterial polymerase differs significantly from the mammalian polymerase, it is possible to develop inhibitors with low toxicity. In addition, because the RNA polymerases of different bacterial genuses have high sequence homology the development of broad-spectrum antibiotics is feasible. Another advantage of this approach is that it targets a metabolic pathway different from those most often affected in cell-based assays, primarily the cell wall and membrane biosynthesis, leading to novel agents with decreased liability for bacterial resistance. Fungal RNA polymerase, similarly, provides a target for antifungal drug discovery.
Traditionally, drug discovery efforts have involved the random screening of natural product extracts or synthetic compound libraries for a desired bioactivity, with medicinal chemistry then being used to optimize the pharmacological and toxicological profiles of the lead compound. With the recent advances in automation and screening technology, the rate at which this process can occur has become limited by compound library availability. In response to this need, a field of chemistry termed combinatorial chemistry has arisen with its objective being molecular diversity generation. The libraries that have been created to date have followed two distinct logical patterns. The first is to mimic the diversity generated in nature by using a limited number of monomeric units (i.e.,. 4 nucleotides, 20 amino acids) and allowing oligomerization to generate the diversity. Originally, the libraries created were either peptides or nucleic acids. However, these types of compounds are usually not acceptable drug candidates due to their poor bioavailability and metabolic susceptibility. The technological advancements afforded by these early combinatorial methods opened the door for the next generation of libraries. The chemistry was developed based on peptide mimetic monomers which when oligomerized became compounds that were not expected to suffer the same biological profiles as their predecessors. Examples of the types of monomeric units that constitute the backbones of these libraries include carbamates, N-substituted glycines, aminimides, and vinylogous amides. While these libraries are certainly useful for in vitro screening, they may, and often do, still possess less than the optimal bioavailability and pharmacokinetic parameters required for oral delivery of drugs. Also, being roughly linear oligomers with limited ability to access the secondary and tertiary structures required for the selective binding observed in natural systems, and not having the thermodynamic advantage of a rigid 3-D structure which is preorganized for the desired binding mode, these library motifs all sample similar molecular space. Consequently, lead compounds isolated from these libraries still require substantial optimization in an ensuing medicinal chemistry program.
Many orally available drugs possess a molecular weight of less than about 500 daltons and routinely contain heterocyclic rings to improve their bioavailability, ease of synthesis, and toxicological properties. These are two of the requirements that are the driving force behind the second major ideological thrust in combinatorial science. Small molecule libraries based on a central molecular scaffold which is subsequently functionalized with pendant groups are currently the subject of a great deal of investigation. There are examples of small molecule libraries that have been prepared by either solid or solution phase chemistry to generate compounds either singly or as mixtures. Examples of the types of molecules that vary around a central core that have been prepared in a combinatorial fashion include benzodiazepines, hydantoins, tetrahydrofurans, xanthines, and cubanes. A distinct advantage to this approach is that a lead structure identified by the screening of these libraries may require less extensive modification in a medicinal chemistry program to produce a drug candidate. Of critical importance in the generation of materials in a combinatorial manner are the questions of coordinate space being searched and diversity of analog functionality. Diversity of physical dimension (X, Y, Z) is important but should be considered in conjunction with molecular properties such as charge, hydrogen bond forming capability, and hydrophobicity/hydrophilicity. The small molecule libraries that currently exist are based either on aromatic scaffolds which impart a planar shape to the analogs prepared, di-functionalized tetrahydrofurans whose derivatives mimic the shape of nucleotides, or poly-functional derivatives of cubane.
The aminocyclitol, 2-deoxystreptamine, spatially resembles a saccharide unit. Molecules that resemble saccharides while lacking the C-1 hemi-acetal found in natural saccharide units, are known competitive inhibitors of a large number of enzymatic systems, especially the glycosidase and glycosyl transferase enzymes. Molecules of this type bind to the active site of the enzyme but are inert to enzymatic transformation. Oligosaccharide portions of glycoproteins and glycolipids located at the cell surface have been associated with such diverse membrane functions as intercellular communication and adhesion of cells, the immune response, and malignancy. The introduction of synthetic sugar analogs into the cell that can regulate the enzymes responsible for biosynthesis of these membrane components has been shown to be an effective method for altering membrane structure and function. However, the number of available non-natural oligosaccharide analogs that can be screened for their inhibitory effects in the cell remains relatively small. This is in part due to the tremendous difficulty associated with oligosaccharide synthesis. The rapid generation of oligosaccharides, and oligosaccharide analogs, is a topic of much current interest. Efforts toward the automated (solid phase) syntheses of oligosaccharides are underway, but this approach has some limitations (such as the inability to prepare anomeric linkages of the .beta.-configuration) and is currently of little utility. In an effort to circumvent the hurdles of oligosaccharide synthesis, a combinatorial approach to oligosaccharide analogs can be conveniently based on 2-deoxystreptamine as disclosed herein.
In addition to the combinatorial generation of possible drug candidates, functionalized derivatives of 2-deoxystreptamine are useful in the field of separation science. Often a very useful source of information for the performance of rational drug discovery is the structure of the biological ligand acceptor. Information obtained by testing combinatorial libraries can be used in conjunction with molecular modelling calculations to simulate modes of binding of the guest ligand to the biological host. The information can then be used to generate hypotheses for structural change of the ligand that are anticipated to enhance affinity.
Unfortunately, the determination of the structures of biological acceptors/receptors or receptor-ligand complexes is an arduous task that requires the purification of the acceptor from the biological milieu from whence it came. Routinely, this purification will entail several types of chromatographic separation as the grail of on-demand biological macromolecule crystallization has yet to be attained.
Chromatographic separations are the result of reversible, differential binding of the components of a mixture to an active surface as the mixture in solution elutes over that surface. Compounds experiencing the greatest associative interaction with the active surface of the solid support will be retained and thereby separated. In the majority of cases, prior art support materials were developed with a specific purpose, or purposes, in mind and are of limited utility. Standard supports used in chromatography, such as reversed phase silica, can denature proteins and often require the use of organic solvents or buffered pH conditions that are incompatible with sensitive biomolecules. When a chromatographic support is derivatized with molecules which bind specifically to a component of a complex mixture, that component will be separated from the rest of the mixture and can be eluted subsequently by an appropriate change in the eluent (ionic strength, pH) that will disrupt the interaction between support and substrate. This so called affinity chromatography is a widely used method for the purification of biological molecules. The development of substances/supports to be used in separations of this type would be very useful.
"Molecular recognition" is a term used to describe the myriad forces that govern the interactions between molecules. Most commonly, this term is applied to the attractive association between two biological macromolecules (e.g., protein-protein, protein-nucleic acid, or nucleic acid-nucleic acid), or to the binding of a small molecule ligand (guest) to a larger macromolecular binding site (host). Often, small molecules can mimic proteins or nucleic acids in their interactions with a binding site and interfere competitively with the macromolecular associations mentioned above. These types of recognition events are at the heart of the drug discovery process. Screening of natural product extracts and random synthetic libraries, for small molecule guests that demonstrate affinity in vitro for a pharmacologically significant macromolecular host has historically been the source of lead molecules in the pharmaceutical industry. Compounds that demonstrate the desired biological activity are then structurally modified, together with what is known about the structure of the binding site, in an effort to optimize their affinity and specificity for the host. This process continues iteratively to generate a structure/activity relationship (SAR) that leads eventually to an optimal drug candidate. This series of events can be performed much more rapidly, and less empirically, if structural data exists for the host or for the natural ligand. Unfortunately, detailed structural information about the biological host is routinely unavailable. Recently, with the advent of automated screening technology, the rate limiting step in the drug discovery process has become the preparation of analogs of an active lead to investigate the structure/activity relationship. The response of the medicinal chemical community to this need has been the active development of a field of chemistry termed combinatorial chemistry.
The various properties of a molecule that can be modified during the course of analog preparation include ionic interactions, hydrogen bonding capability, hydrophobicity, .pi. stacking interactions, Van der Waals or steric interactions with the host, and reorganization of the molecule into a conformation resembling the bound conformation a natural ligand.